Baseten vs Replicate
Same workloads, both price lists, refreshed daily. On shared line items today: Baseten is cheaper on 5 of 6 shared models (input or output price), Replicate on 0.
Baseten
Production inference platform with Truss packaging, optimized serving engines and enterprise-grade autoscaling. Strong for teams shipping custom models with SLAs.
- Optimized model serving (TensorRT-LLM)
- Enterprise autoscaling and observability
Replicate
Run thousands of community models (LLMs, image, audio, video) behind one API with per-second GPU billing: the fastest path from 'found a model on GitHub' to production endpoint..
- Massive community model zoo
- Per-second billing, scale to zero
- Custom models via Cog
Shared models, priced by both
List prices refreshed 2026-11-22 ยท input $/1M tokens| Model | Baseten in/out | Replicate in/out | Cheaper input |
|---|---|---|---|
| GPT-OSS 120B | $0.100 / $0.500 | $0.180 / $0.720 | Baseten |
| DeepSeek-V3.1 | $0.500 / $1.50 | $0.672 / $2.02 | Baseten |
| Kimi K2.5 | $0.600 / $3.00 | $0.720 / $3.60 | Baseten |
| Kimi K2 Thinking | $0.600 / $2.50 | $0.660 / $2.70 | Baseten |
| DeepSeek V3 | $0.770 / $0.770 | $1.45 / $1.45 | Baseten |
| Kimi K2.6 | $0.950 / $4.00 | $0.950 / $4.00 | tie |
Cheaper-on-count is a tally of listed prices, not a quality verdict: throughput, reliability and quantization differ between providers. Disclosure.